Hard Landings

When your assignment is to put a space probe on another planet, be prepared to sweat.

Viking lander footpad pia00381.jpg
The first picture taken by Viking 1 on the surface of Mars, July 20, 1976.

Five minutes. That's how long it will take for Mars Pathfinder to make this Fourth of July either a day of celebration or a day of sorrow for Tony Spear. Just five minutes after speeding into the Martian atmosphere at 17,000 mph, the Pathfinder lander will strike the Red Planet. If it survives, this robotic emissary, which was launched by a Delta rocket last December, will undertake the first exploration of Mars' surface in more than two decades. If it does not, Spear and his team of engineers and scientists--who labored for more than five years to create Pathfinder and send it to Mars--will shoulder one of NASA's greatest disappointments. Of those five minutes, Spear says: "I don't know how I'm going to stand it."

But Spear is only the latest in a long line of engineers to face the uncertainty that precedes every robotic landing on another world. It's an anxiety that began almost four decades ago, when the cold war was being waged in space. In the years after Sputnik, with the United States stinging from one Soviet space first after another, engineers at NASA's Jet Propulsion Laboratory in Pasadena, California, dreamed of scoring a first of their own. Although the Soviets had, in September 1959, already hit the moon with their Luna 2 probe, a true moon landing had yet to be achieved. That was the goal of JPL director William Pickering, who pushed for a mission to deliver a package of scientific instruments to the lunar surface. Early in 1960, NASA headquarters gave formal approval to JPL's Project Ranger.

In terms of physics, the concept for any lander is simple to describe: Just as it reaches its destination, it must undo the work of the launch rocket, canceling the kinetic energy the launch and the gravitational pull its destination give it. But Ranger project manager James Burke and his team found that executing this task was anything but simple. By the time Ranger reached the vicinity of the moon it would be traveling at 4,500 mph. Ideally the lander would settle onto the moon at only a few miles per hour. But this so-called "soft landing" was beyond the reach of existing technology. The best Ranger could do was fire a blast from a solid-fuel braking rocket to slow its descent before its lander simply fell to the surface--a "hard landing."

Solid-fuel rockets were already being used in military airlifts to supplement parachutes when tanks and other massive objects were dropped out of airplanes. But the JPL engineers knew that solid rockets, while simpler than liquid-fuel ones, were also more unpredictable. No one could be sure a solid rocket would deliver the amount of braking needed to counteract all of the lander's excess speed. Furthermore, nobody knew how to predict precisely how fast the spacecraft would be going relative to the moon, or even the exact location of the moon itself.

With all these uncertainties, Burke figured that Ranger might strike the moon at speeds up to 200 mph. He and his team began talking about developing a rugged spherical capsule capable of withstanding such an impact. If this "survival package" seemed a less than elegant plan for humanity's first landing on the moon, Burke didn't mind. "All we were thinking about," he says, "was 'Let's get a transmitter down so we can prove we're there.' "

But how to protect sensitive scientific instruments from a crash as violent as an Indy race car hitting a concrete wall? To identify the best energy absorber, a variety of materials, including aluminum honeycomb, cardboard, and, in Burke's words, "anything crushable," were subjected to tests such as being dropped from a helicopter and slammed around with laboratory equipment. The victor, by a surprisingly wide margin, proved to be blocks of balsa wood, oriented with the end grain radiating out around the sphere for maximum energy absorption.

By the summer of 1960, a 26-inch-diameter sphere weighing 92 pounds began to take shape at a division of the Ford Motor Company in Newport Beach, California. Attached to the capsule would be a solid-fuel retrorocket, which was to ignite when Ranger was 10 miles above the moon. Ten seconds later, after slowing the lander almost to a hover, the rocket would burn out and be cast off. Pulled by the moon's gentle gravity, the sphere would fall the remaining 1,100 feet to the surface, striking at a speed of about 75 mph. Cushioned by a six-inch layer of balsa wood, the lander would bump and roll to a stop. Inside, floating on a thin layer of water, a one-foot-diameter fiberglass sphere containing a seismometer, radio, and batteries would right itself and begin transmitting.

But the lander was never able to prove it was up to the task. Rangers 1 through 6, including the only three to carry survival packages, all failed. Several missed the moon because of malfunctions in the Atlas Agena launcher. Launch calculations showed that Ranger 4 reached the lunar surface, but the spacecraft malfunctioned shortly after leaving Earth in April 1962 and was unable to return data. Although Rangers 7 through 9 were successful, taking high-resolution photographs on the way down to a crash-landing on the moon, none carried survival packages--in part because JPL was by then working on a lunar lander called Surveyor.

Looking back, Burke sees the Ranger lander as a challenge undertaken before its time. "We were trying something that was too complicated," he says. "Our reach was exceeding our grasp."

But Surveyor made Ranger look simple; it would have to execute the first soft landing. Surveyor would build on Ranger's braking technology, using a solid-fuel rocket to slow from an initial descent at more than 6,000 mph to 240 mph. However, the lander would have to aim the thrust of its braking rocket directly along its flight path to avoid tumbling. The task was daunting. "We were starting essentially from scratch," says engineer Leo Stoolman, who managed Surveyor's design and construction at Hughes Aircraft in El Segundo, California. When the performance of the new Atlas-Centaur launch vehicle fell short of predictions, design teams were forced to trim hundreds of pounds from Surveyor's allowable weight. More than 60 percent of the spacecraft's final 2,200-pound weight went to the braking rocket; what remained was barely enough to carry out a scientific mission after landing.

Surveyor's braking rocket was so heavy because it had to deliver up to 10,000 pounds of thrust, the cost of counteracting energy rather than absorbing it, as Ranger did. To guarantee that the rocket's thrust travel as precisely through Surveyor's center of gravity as possible, engineers tried to make the rocket's nozzle perfectly symmetrical and aligned with the rocket's center of gravity. "It looked like it ought to be doable," Stoolman remembers, but there was no way to simulate accurately enough to be sure.

Then, at about 25,000 feet above the moon's surface, Surveyor would use small, liquid-fuel rockets to slow to about 8 mph in a vertical descent. The spacecraft would rely on these rockets, called vernier engines, for both control and descent the rest of the trip down. Their control system would use signals from a radar altimeter to measure distance from the surface and a Doppler radar to measure speed. The radars, which used new technology, also caused their share of headaches. And the verniers' development problems were so distressing that in 1964 JPL canceled its contract with Thiokol Chemical's Reaction Motors Division--only to re-award it a short time later, after the company continued working on the engines and made promising advances. With so many problems, says Stoolman, "we were running scared all the time. We didn't take anything for granted."

Technical burdens aside, one unknown loomed larger than any other: the nature of the moon's surface (see "A Smooth Spot in Tranquility," June/July 1989). Faced with uncertainty, the Surveyor teams designed the lander for a range of conditions. To absorb an impact on hard ground the three landing legs were fitted with shock absorbers, the footpads were made of crushable aluminum honeycomb, and three more honeycomb blocks were mounted underneath the craft's spindly frame. If, on the other hand, Surveyor encountered a dust layer several feet thick, as one scientist predicted, the engineers hoped its oversize footpads and wide body might act like snowshoes to keep it from sinking too far.

The Surveyor teams weren't the only ones grappling with unknowns. By this time NASA had been directed to put a man on the moon; Ranger and Surveyor were officially Apollo's advance team. But there were times when the NASA engineers working on Apollo seemed uninterested in Surveyor or even disdainful, a reflection of the rivalry between the space agency's manned and unmanned programs. Late in 1961 a Surveyor team visited NASA's Langley Research Center in Virginia to ask what their spacecraft could do to help Apollo. One Langley engineer responded, "Crash into the moon and smash all to hell." At least then, he added, they would know the surface was solid.

Surveyor engineers may have recalled that comment in April 1964, when a test version of the spacecraft, outfitted with working radars and vernier engines, was suspended from a balloon and lofted 1,500 feet above the New Mexico desert for a test of the landing system. Before the test could begin, a nearby electrical storm triggered the balloon's electronic release mechanism and the test Surveyor fell to the desert floor and broke into pieces.

For JPL, Surveyor's troubles worsened an already dire situation. Ranger 6 had failed that January. The bad news persisted even after Ranger 7 finally triumphed in July: A second Surveyor drop test failed in October because of a series of malfunctions. When engineer Robert Parks took over as project manager of the Surveyor effort, his colleagues at the lab offered condolences. "They thought I had taken on an impossible task," he says.

The delays cost the Surveyor team the moon race. On February 4, 1966, a Soviet soft-lander called Luna 9 alighted on the plains of the moon's Ocean of Storms and radioed back pictures. Luna didn't match Surveyor's technical sophistication, but that didn't lessen the sting at JPL. NASA's own attempt was only four months away.

On June 1, less than three days after a flawless launch, Surveyor 1 reached the vicinity of the moon. In Pasadena it was past 11 p.m., but JPL was alive with activity. Inside the lab's new mission control building, Parks and his team knew that everything in the landing sequence would happen automatically; all they could do was wait. Less than 50 miles above the moon, Surveyor 1's braking rocket fired a 40-second blast, then fell away. To everyone's relief, telemetry showed the lander wasn't tumbling. At 25,000 feet the verniers took over.

A mission commentator called out the diminishing altitude: 1,000 feet, then 500, then 50, then 12--and finally "Touchdown."

In mission control, no one could believe it--well, almost no one. "My feeling," says engineer Gene Giberson, "was, 'Yep, we did it, we did exactly what we said we were going to do.' " Geologist Gene Shoemaker, a Ranger veteran leading one of the Surveyor science teams, recalls having a different reaction: "My God! It landed!

"We were all shell-shocked from Ranger," Shoemaker says. "Hell, I wouldn't have given you a 10 percent chance that Surveyor 1 was going to land." Not only had Surveyor 1 landed on the Ocean of Storms, it had done so at a comfortable 10 mph. Half an hour later, the first television images began to appear on the monitors at JPL, showing a round footpad perched on a dusty but firm moonscape. As Surveyor's pictures revealed a 100-foot crater rimmed with boulders, it became clear that getting Surveyor down safely had taken more than ingenuity. "I think we were all aware that it was going to be a matter of luck," Shoemaker says.

For the most part, Surveyor's luck held. Six more missions followed; all but two were successful. For the program's climax in January 1968, Surveyor 7 made the riskiest landing of all, touching down in the rugged highlands next to the giant crater Tycho. Eighteen months later, when the Apollo 11 astronauts descended to the Sea of Tranquillity, Neil Armstrong had to take over manual control from the onboard computer, which was aiming for a giant, boulder-strewn crater. Today Gene Shoemaker marvels at the way things turned out. "Every time we landed blind with Surveyor and had a chance to land, we landed successfully," he says. "As luck would have it, you really had to have the astronaut there to land the sucker on [Apollo] 11. Surveyor would've been a gone gosling."

By the time of Apollo 11's success, NASA was planning to attempt a landing far beyond any astronaut's reach. Since 1962 the agency had been studying how to put a spacecraft on Mars. Everything about the idea overshadowed Surveyor's challenges, beginning with Mars' vast distance. The trip from Earth would span 10 months and more than 200 million miles, requiring a more self-reliant and more reliable spacecraft. Compared with all previous landers, any Mars lander worth sending would be not only more complex but heavier, and would necessitate a more powerful booster. The Martian gravity, about three times that of the moon, required the descending probe to withstand greater forces of deceleration. But to scientists, the Red Planet's pull was irresistible.

At the Langley Research Center, the effort was led by engineer Jim Martin, whose formidable presence and military-style crewcut won him nicknames like "the Prussian General." Viking needed a tough manager; in addition to being the most challenging robotic mission NASA had ever planned, the Mars landing was the most expensive.

Although the Viking landers were similar to their lunar predecessors, Martin's teams looked to the Apollo lunar module rather than Surveyor for technological hand-me-downs. Viking, like Apollo, would also use an orbiter from which the lander would be dispatched. Mars, however, has something the moon lacks: an atmosphere. From the Mariner Mars flyby probes, which had begun in the mid-1960s, scientists knew that the planet's carbon dioxide envelope was tenuous, with a surface pressure only a small fraction of that on Earth. Still, it was enough to require that Viking use a heat shield, followed by a parachute to decelerate. The tried-and-true combination of radar and vernier rockets would take care of the rest.

Even more troublesome than getting to the surface of Mars was keeping stowaways from going along. A spacecraft created to detect Martian microbes would have to be sterilized before leaving Earth. For Viking, that meant baking the entire lander for 40 hours at temperatures up to 234 degrees Fahrenheit. That played havoc with microelectronic components--and both landers were full of them, including a miniature biological laboratory in each. Just as vulnerable were the twin onboard computers, each possessing 18,000 words of memory in a container roughly the size of an overnight bag--a marvel by early-1970s standards. The entire descent to Mars would be controlled by one of those computers, but sterilization almost did them in; in one heat test after another the magnetic-wire memory failed. The computer occupied a spot on Jim Martin's infamous Top Ten List of problems for two and a half years, until engineers at Honeywell Aerospace finally perfected it.

In the end, Martin estimates, sterilization soaked up a quarter of the $930 million that the four Viking spacecraft cost all together--about $3 billion in current dollars. "They wanted the best," he says of NASA headquarters.

Even so, some inside the project feared that all that money, time, and effort might be for naught. The odds weren't promising. Not long after the Vikings left Earth in August and September of 1975, project scientist Gerry Soffen asked chief engineer Israel Taback what he thought the chances were of one lander getting down safely. When Taback estimated the odds at only 30 to 40 percent, Soffen says, "I was surprised they were that good!" He fully expected that Viking 1 would crash; then everyone would try to figure out what went wrong in time for Viking 2's attempt weeks later.

Jim Martin was more optimistic--but not much more. During a meeting with Viking scientists the following spring just before the spacecraft's encounter with Mars, Martin had said the odds of success were 50 percent. One scientist threw up his hands and wondered aloud why he had invested a decade of his life in a project with so little chance of succeeding. Martin explained that this was why there were two sets of spacecraft.

The Soviets knew the risks all too well. By 1976 they had tried as many as six times to place a lander on Mars without real success. In 1971 the Mars 3 lander reached the surface and transmitted for 20 seconds, then mysteriously died. Two years later Mars 7 missed the planet by 800 miles because of an internal malfunction; three days after that, Mars 6 crashed on the surface--and gave Viking the chance to make history.

On June 19, 1976, Viking 1 went into orbit around Mars. Three days later, when the orbiter's first images were received at JPL, scientists were stunned--not only because the pictures showed more detail than any previous views of Mars, but because the planned landing site was far rougher than anyone had expected. So began a feverish effort to find a safe landing site, using the Viking 1 orbiter's cameras. Graduate students poured over each new image with magnifiers, tallying the smallest boulders and craters for hazard analysis. Scientists argued about how to interpret radar soundings made with Earth-based radio telescopes.

Viking's earliest possible landing date, by coincidence, had turned out to be July 4. Now NASA was counting on the landing to be part of the nation's bicentennial celebration. But as the hunt for a new landing site dragged on, Martin informed NASA administrator James Fletcher that the July 4th date was out of the question. By chance, Viking 1 was ready to land on a day that had its own significance: July 20, seven years to the day after the first humans had landed on the moon.

The moment of truth came just after 5 a.m. Pacific time, as the lander, encased in its heat-protective aeroshell, sped into the Martian atmosphere. Within minutes deceleration forces on the lander built to 8.4 Gs, then slacked off somewhat as the aeroshell sped toward Mars' Chryse Plains. At 21,000 feet, with Viking still traveling at supersonic speed, a mortar fired to deploy the parachute. Less than a minute later the parachute was cast off as the lander's three vernier engines ignited.

Like their Surveyor predecessors, Jim Martin and his people were reduced to waiting. Because Viking's radio signals took 19 minutes to reach Earth, it was a bit like sweating out the delayed broadcast of a crucial football game; by the time the first descent data appeared on their screens they knew that Viking would have already landed--or crashed. Then came exultant words: "Touchdown! We have touchdown!"

When the first image from the surface of Mars appeared on the monitors, NASA knew it had been lucky again: A boulder almost as big as the lander stood only 30 feet away. Less than seven weeks later, Viking 2 alighted safely on Mars' Utopia Plains, but instead of the sand dunes the geologists had promised, its pictures showed a jumble of rocks, much to Martin's chagrin. But no matter: The Mars landings had been more successful than almost anyone had dared hope.

Even so, another 16 years would pass before NASA decided to return to the surface of Mars. Not until 1992 did the space agency approve Mars Pathfinder as the first of a series of robotic landings. As project manager Tony Spear soon learned, Pathfinder was expected to live up to its name and test a new type of landing system. Spear's assignment got even tougher when he was given a budget ceiling of $150 million in 1992 dollars--five percent of Viking's cost. Says Spear, "We had no idea whether we could do that or not."

The Pathfinder used a risky new method of arrival -- bouncing to a halt inside a protective cocoon of airbags.

Whatever they lacked in certainty Spear's people made up for in motivation. Pathfinder represents a new generation of scientists and engineers, exemplified by 36-year-old Sam Thurman. An engineer who coordinated the design of Pathfinder's entry, descent, and landing systems, Thurman remembers being galvanized by the Viking landings as a teenager. When he came to JPL 10 years ago, nothing like a Mars lander was in the works. "I was really getting worried for a while that the most exciting era in the space program had come and gone, and I missed it," Thurman says. Pathfinder was the opportunity he had been waiting for.

Thurman and his teammates had only 38 months to redefine the art of landing on Mars, and they would have to do it with a fraction of Viking's workforce. They also had to stay motivated in the face of some healthy criticism from a NASA review panel that included Viking veterans like Jim Martin. (Engineers who have already braved the challenges of landing on another world can be tough to impress. At a 20th anniversary reunion last summer at the National Air and Space Museum, one Viking veteran expressed doubt that the public would be excited about Pathfinder or its rover, which he puckishly dismissed as "a Tonka toy." Gesturing toward a Viking replica, he beamed with pride. "Viking was big! That was a real lander.") "They were brutal," Pathfinder deputy project manager Brian Muirhead recalls wryly. "They were sharp guys and they were getting paid to beat us up. I think they sensed that it was for our benefit."

By 1993, when work on Pathfinder began, the reviewers had told Spear and his team that the only way to avoid smashing through their cost ceiling was to simplify as much as possible. There would be no orbiter; like Surveyor, Pathfinder would fly directly to the surface. (An orbiter launched by NASA last November, Mars Global Surveyor, will reach Mars in September to provide a detailed look at the planet from orbit. But by then Spears hopes Pathfinder will be part of the Martian scenery.)

As much as possible, the design of the lander's heat shield and parachute are borrowed from Viking. But the landing system is entirely different: There is no retrorocket and no radar-controlled vernier engines. Instead, Spear's team chose an approach reminiscent of Ranger's survival package. Pathfinder will simply fall onto the surface, protected not by balsa wood but by a set of pressurized airbags. Airbags had been studied for space probes by JPL engineers as early as 1966. Now, Spear hoped, they would be not only the most rugged solution, but the cheapest.

Like every landing project before them, the Pathfinder team found that nothing was easy. Yes, technology had come a long way since Viking, and Pathfinder's computers and electronic components were cheaper to build. Then, too, the requirement for pre-launch sterilization was dropped, because Viking data suggested that the Martian surface is not only lifeless but chemically hostile to organic matter. And although the Viking heat shield and parachute designs had to be modified for Pathfinder, Spear says using them "saved us a ton of money." But hopes of building Pathfinder without a retrorocket proved too optimistic; without one, calculations showed, the lander would hit the surface too fast for airbags to protect it. A trio of small solid-fuel rockets had to be added, along with a simple radar altimeter to signal when they should ignite.

Meanwhile, work went ahead on the airbags, which turned out to be the biggest headache of all. By the end of 1994 a group led by JPL's Tom Rivellini had worked out their basic design. The spacecraft would be shaped like a tetrahedron, with three sides that opened like flower petals to let it right itself after landing. Each petal would contain an airbag with six spherical lobes that would be deployed just before landing. Fully inflated, the airbags would make Pathfinder resemble a bunch of party balloons. Of course, they would have to be enormously more rugged; calculations showed that the 800-pound lander could strike the surface at 60 mph.

Rivellini's group had selected a material they felt was up to the task: a fabric called Vectran, an exotic cousin of the more familiar Kevlar used in bulletproof vests. To fabricate the bags, they chose the Delaware-based ILC Corporation, whose products include NASA spacesuits. But it didn't take long for frustration to set in. Vectran's incredible resistance to stretching--the source of its strength--made it extremely intolerant of any errors in the shape of a seam. The first bags stitched together at ILC broke apart when they were pressurized.

Even as that problem was solved, it became apparent that Vectran had another weakness, this one potentially fatal. Although tests showed that a Vectran airbag would resist being punctured if a sharp rock were pushed straight into it, a rock dragged across the fabric would disrupt the weave of the fibers, making the bag dramatically more vulnerable to tearing. For that reason, it was crucial that Pathfinder strike the surface with as little horizontal motion as possible. The problem was, the Martian winds might not cooperate. Wind gusts hadn't been a problem for Viking, whose vernier engines were able to counteract them. But Pathfinder would make its final descent suspended from a parachute and aeroshell at the end of a 65-foot tether called a bridle. Blowing winds could make the lander swing like a pendulum, giving it horizontal speed when the bridle is cut. Even worse, a gust could tip the aeroshell at the moment the retrorockets fire, propelling Pathfinder toward the ground at a shallow angle.

Fortunately, there was a tool that no previous landing project had available. Sam Thurman had created a computer program capable of simulating the entire landing sequence. It showed that the parachute-bridle-lander combination would respond to winds not by swinging from side to side but by vibrating like a plucked guitar string, with the lander nearly unaffected. And if the aeroshell tipped, Thurman found, it would quickly right itself.

Still, the risk of horizontal speed at impact continued to cause worry. Says JPL's Rob Manning, the engineer in charge of getting Pathfinder down safely, "We realized we didn't want a regular airbag--we wanted a Michelin tire." Or at least something as strong as a steel-belted radial; the airbags also had to be small enough to be packed with the lander inside its aeroshell, and light enough not to exceed Pathfinder's stringent weight limits.

In late 1994, Rivellini's group began conducting drop tests of the bags with a model lander. To simulate Martian atmospheric surface pressure--only 0.7 percent that on Earth--Rivellini's group turned to the largest vacuum chamber in the country, near Sandusky, Ohio, operated by NASA's Lewis Research Center. Inside the cavernous chamber, a model Pathfinder was suspended 80 feet above the floor, connected to a set of giant bungee cords that would allow it to plunge downward until it crashed into a tilted platform studded with sharp rocks.

Just as Rivellini had feared, the rocks tore the airbags, even though the bags had been covered with a second layer of material to resist abrasion. He still sounds amazed when he talks about how badly the Vectran fared. After one run at speeds approaching those of actual touchdown, "it looked like a bear walked up to this thing with its claws and just shredded the hell out of it," Rivellini recalls.

Fortunately, ILC seamstress Eleanor Foraker--who had stitched the moon suits worn by Neil Armstrong and Buzz Aldrin--was on hand to make repairs at a momentÕs notice. That was no easy task; Foraker had to use special ceramic shears for the job. As Rob Manning explains: "You can't use scissors on this stuff." And fabric designed not to be penetrated was "godawful to stitch," Rivellini adds.

The failures continued into the summer of 1995, with time running out to build and test the lander's components. No computer simulation could solve the airbag problems; Rivellini's group had to rely on trial and error as they experimented with different thicknesses and densities of fabric. Finally they hit on a breakthrough: Multiple layers of lightweight fabric worked better than a single heavy layer.

Even as this hopeful sign emerged, a slew of other airbag-related problems demanded solutions. How would the bags be packed into the tiny space between the lander and its aeroshell? At ILC, a former Air Force paratrooper devised a folding scheme. What about inflating them quickly and safely? That called for hot-gas generators, which burn the same type of propellant used for the retrorockets. Rivellini's group knew Pathfinder would land during the Martian nighttime, with temperatures hovering at 112 degrees below zero. As the gas inside the bags chilled, they would lose pressure in moments. To prevent this, a second set of gas generators was added. Of course, the bags would have to be deflated after landing. And finally, Rivellini had to solve a problem that some at JPL predicted would be Pathfinder's undoing: how to retract the deflated bags so that the lander could open its petals and begin operating. After a series of low-tech experiments ("trash bags," Rivellini explains) he devised a set of Vectran cords to be routed through loops attached to the bags' inner walls; after touchdown the cords will be pulled into the lander using a winch.

By the end of 1995 everything had come together. Rivellini and his colleagues had created a bag that didn't break, with up to four layers of Vectran in the most vulnerable places. In August 1996 Pathfinder was shipped to Cape Canaveral to be prepared for its December launch. Its next stop would be the mouth of the Martian channel called Ares Valley, the very place Viking 1 was originally slated to land. All the ingredients seemed to be in place for a bold and ambitious landing, if that is what the space agency is looking for. But as Tony Spear recalls, one high-level NASA manager seemed to have mixed feelings about that: "He told me, 'Don't you dare fail. If you do, I'll shoot you on the [JPL] mall.' "

Jim Martin, who understands that kind of pressure, nevertheless has doubts about whether Pathfinder will succeed. "If I had to put a probability on it," he says, "I'd have a problem." Much of that familiar uncertainty, Martin says, could be avoided if future landers had some kind of hazard avoidance system, borrowing from the technology now used in "smart" bombs and missiles. It would be expensive, Martin realizes, but "having a failed lander mission is not cheap either."

There is much at stake beyond Mars Pathfinder. In 1998 NASA will launch a lander resembling a smaller version of Viking to set down near the ice cap at the planet's south pole. And as early as 2005 will come the Holy Grail of planetary exploration, a mission to retrieve a sample of Martian rock and soil. Now that scientists may have discovered evidence of fossil life inside a meteorite from Mars, there is more interest than ever in a sample-return mission, an engineering task that presents its own knotty problems. The sample-return lander may use some combination of the techniques worked out for Pathfinder and the Mars '98 lander.

But Tony Spear won't be thinking about any of that on July 4, as Pathfinder comes screaming into the Martian atmosphere. Spear, like everyone involved, knows that in many ways Pathfinder is the most complex lander yet devised. After all the testing his team has done, he says, he is confident that the fast-paced events in PathfinderÕs descent will all go off without a hitch, as they must. But when the time comes, they will still seem to take forever, as Pathfinder endures the heat of reentry, deploys its parachute, and lowers itself to the end of its bridle.

Just eight seconds away from impact, the airbags will inflate--explosively, in an instant. Four seconds later the retrorockets will ignite, lighting up the Martian night with their own Fourth of July fireworks, to slow the lander to a halt a little less than 50 feet above the ground. Severing the bridle, Pathfinder will fall onto Mars, bouncing for perhaps two minutes before coming to a stop, its 200 million-mile journey finally over. Just as their predecessors did 21 years before, Spear's team will have to wait long minutes for the radio signals to travel to Earth before they know if they were successful. Hours later, as the sun rises over Ares Valley, Pathfinder will come to life and deploy a 22-pound rover called Sojourner, and the exploration of Mars will begin again.

Will it work? "Yes," says Spear without equivocation. "But I'm still scared to death."

Get the latest stories in your inbox every weekday.